Message Order

[Tim Reid]

Richard A. Franklin wrote:
> Jonathan Hutchins wrote:
> 
>>On Tuesday 18 November 2003 08:09 am, Dave Hull wrote:
>>
>>  
>>
>>>Why are computer clocks so good at gaining/losing time?
>>>    
>>>
>>
>>There's gotta be a FAQ on that somewhere, but basically it's because they were 
>>never designed to keep absolute time, just "relative" time.  They provide the 
>>metronome beat that the system runs by and make it possible to tell a "new" 
>>file from an "older" file.  The original P's didn't even store time when you 
>>shut them off, you had to set it at each startup.  There were piggyback chips 
>>you could put over the BIOS that had an internal clock and battery, and 
>>that's when people started expecting their PC to know what time it really 
>>was.
>>  
>>
> http://www.linux.se/doc/HOWTO/mini/Clock-1.html
> 
> The info on this page seemed relevant.

IIRC ( I can't find the webpage).  I read something about the processor 
clock drifting because of the various different timings needed for 
different tasks.  Then, when the computer halts, most OS's sync the BIOS 
clock to the processor clock.  Then the BIOS time is off.  Have you 
noticed that a computer that is up 24/7 without time correction has 
usually drifted much further than a computer that is usually down/off 
most of the time?  This is my experiance with PC's anyway.  The place 
where I work uses a DOS based system to keep track of employee shifts, 
and I've been a few minutes late because the system is sometimes as much 
as 10 minutes fast.